Agentic Analytics in Finance: Lessons from Navan and EcoLab

Finance leaders are operating in one of the most demanding macro environments in recent memory.  Interest rates are moving faster than most models anticipated, reshaping the cost of capital almost overnight. Supply chain fragility has also turned working capital management into a moving target, and geopolitical uncertainty is changing how you plan for the future. 

Yet for many finance functions, the analytics stack hasn't kept pace with that urgency. 

The dashboards you used to rely on were built for a world where finance had weeks to prepare management reporting. They were designed for questions asked last quarter, not the ones landing in your inbox this morning.

Meanwhile, many AI platforms have added another layer of complexity: impressive demos, unclear governance, and very little operational trust. For finance leaders who need answers they can act on—and defend—that's not good enough.

Here’s how leading organizations like Ecolab and Navan are solving for these challenges in 2026.

How Ecolab Solved the BI Backlog Problem

Every CFO has lived this: a critical question arrives from the CEO, the board, or a business unit leader, and the answer sits somewhere in the data. Getting to it means submitting a request, waiting for an analyst to build a report, reconciling numbers across systems, and eventually surfacing an insight that's already 48 hours stale.

This is an architecture failure. Most analytics platforms were designed to serve pre-defined questions through pre-built dashboards; the moment a question falls outside that scope, the bottleneck appears.

Ecolab, the global sustainability and hygiene leader, knew this problem well. Their month-end close was Excel-heavy, disconnected, and dependent on individuals manually hunting down KPIs to get numbers into a deck by day five. Key metrics weren't live operational data. Someone had to pull, reconcile, and package them first. 

The finance data team had become, in effect, a reporting service rather than a strategic function.

Ecolab: the Impact of an Agentic System

The solution went deeper than a tool change. Ecolab consolidated finance KPIs into a shared data model that worked across ERPs and regions, preserving transaction-level detail instead of forcing finance leaders into static summaries. Metrics that once required tracking down a person now arrive automatically, through alerts and notifications, before the question is even asked.

That shift—from reactive reporting to proactive intelligence—is exactly what the next generation of financial analytics is built around.

How Navan Designed Agentic Analytics for Trust

Finance is a domain where the stakes of a wrong number are unusually high. Inconsistent numbers don’t just slow decision-making, they erode confidence in the finance function itself. In finance, the intelligence layer is only as trustworthy as the data foundation beneath it.

Navan, the travel and expense management platform, approached this as a systems design problem rather than an AI rollout. 

Navan the Impact of an Agentic System

Snowflake provided the data foundation, dbt handled metric transformations, and ThoughtSpot’s semantic layer created shared, governed business definitions across the organisation. 

On top of that, the Spotter agent gave business users a conversational way to explore finance data in natural language, without bypassing the controls underneath. 

The result? A finance analytics environment where speed and trust are built into the same system.

When the data foundation is governed and the definitions are shared, speed stops being the enemy of accuracy.

What Does Agentic Analytics Look Like in Finance?

Traditional BI systems were designed to answer predefined questions. Agentic analytics has changed that model in a few ways:

1. It Navigates Data Autonomously

Rather than requiring a user to know which table, which filter, or which pre-built chart to consult, an agentic system takes a plain-language question—"which customers have invoices more than 60 days overdue, and what's our total exposure?"—and constructs the analysis across the relevant data domains, without human intervention in the query logic.

1. It Maintains Context Across a Conversation

A single question rarely tells the full story. Agentic analytics preserves the thread so when a CFO asks "show me cumulative revenue by month" and follows up with "now break that down by region and compare to last year," the system remembers the context of the conversation and builds progressively, the way a skilled human analyst would.

1. It Shifts the Finance Team's Role

When business leaders can self-serve trusted answers, the finance team moves out of the report factory and into a genuinely strategic role—one focused on modeling and judgment, not request fulfillment. We'll see what that looks like at organizations already running this way.

What Does Agentic Finance Analytics Look Like in Practice?

Consider a typical morning for a VP of FP&A. Before a leadership call, they need to understand three things: where order intake is tracking against plan, whether working capital is within acceptable bounds, and which customers are creating AR risk.

In a traditional analytics environment, that's three separate reports, two different systems, and a call to an analyst for the AR aging cut.

In an agentic environment, the conversation looks like this:

• "What is our total order value and backlog this quarter, by region?"- answered in seconds, with backlog calculated dynamically based on open orders not yet shipped.

• "Show me DSO and DPO trends for the last six months.” - with outliers and regional shifts flagged automatically.

• "Which customers have overdue invoices beyond 60 days and what's the dollar exposure?" - returning a ranked list of at-risk accounts and total exposure.

This is agentic analytics for finance.

Finance teams building on the Spotter agent are doing this today—asking questions of their data in natural language, drilling from summary to transaction level, and sharing analysis directly with stakeholders.

Your Architecture Determines Your Outcome

Interest rate uncertainty, currency volatility, evolving regulatory requirements, and continued pressure to do more with less will keep demands on the finance function elevated for the foreseeable future.

See how ThoughtSpot Spotter helps finance teams explore governed data and move from reactive reporting to proactive decision-making. Book a demo today.